Six stores, one auction: how regional marketing makes same-brand dealers cannibalize each other

Short answer: nobody in the program is doing anything wrong, and the result is still that the market buys the same shopper six times. When a regional office is measured on participation and share of voice, and a store is measured on cost per unit sold, both scorecards can turn green while the metro's registrations do not move. That is the shape of the problem. It is structural, not personal, which is also why replacing the agency never fixes it.

Two scorecards, one auction

A regional marketing manager answers to a participation percentage, a brand share target, a co-op utilization number and a regional review deck. A dealer answers to units sold this month and the cost of getting them. Those are not the same objective, and they rarely point the same direction.

The regional office cannot tell six stores in one metro to stop competing, because state franchise law does not let an OEM dictate dealer advertising prices or impose exclusive territories at will. What an OEM does control is co-op eligibility: which vendors, which campaign structures, which creative, which keywords qualify for reimbursement. That makes funding the lever. And a lever that rewards every dealer for running the approved campaigns will produce exactly the outcome it pays for.

What cannibalization looks like when nobody is trying

Four mechanics, each individually defensible, stacking into the same result.

Duplicate coverage. Six rooftops in one metro, each running the same four campaign types against the same trade areas. Overlap is not an edge case, it is the default geometry of a metro.

Brand-term bidding. Every store wants its name and the brand name. So does the regional office, at the top of the page. The same auction now includes the brand buying against itself.

Co-op eligibility as pressure. A store that declines the approved program loses funding it paid into. A store that takes it runs the campaigns the vendor built for the average store, not for the market it is in.

A share-of-voice scorecard. Share of voice is a ratio the brand can improve by spending more into the same demand. Nothing in that metric requires the spend to be incremental.

The arithmetic of buying the same shopper six times

Take one in-market shopper researching a model. Six stores plus the brand can each reach them. The first impression is doing real work: it creates awareness of a specific store, an offer, a reason to act. The second adds some reinforcement. By the sixth, the marginal impression is close to decorative.

Its cost is not decorative. All six bids sit in the same auction, so the price of reaching that shopper is set by the highest intent to reach them, which is the neighbor with an aged unit and a month-end number. Duplicate coverage does not split the cost of demand. It raises it, then divides the outcome.

The signature is boringly consistent:

• Market branded spend rises quarter over quarter.
• Brand-term cost per click climbs.
• Market registrations do not move.
• Every individual store reports decent cost per lead, because leads are cheap and attribution is local.

That last line is why the problem survives reviews. Cost per lead is a store-level ratio that flatters duplicate spend, because a shopper who sees six stores is six leads. Cost per incremental unit is a market-level ratio, and almost no co-op program computes it.

The regional office is not being stupid

It is being asked to hit a number it can only move with the tools it has. It cannot set dealer prices. It cannot hand one store a metro. It cannot audit every rooftop's auction overlap without data most dealer groups do not consolidate. So it does the only thing available: it funds more participation and reports the participation rate, then reports share of voice as evidence that the market is being worked.

The honest version of the criticism is narrower than it usually sounds. The regional office is not wasting money. It is spending money on demand capture and calling it demand creation, and it cannot see the difference because nobody hands it registrations by trade area.

The split that actually works

Draw one line and hold it: the brand funds demand creation, stores fund demand capture.

Demand creation belongs to the brand, because no single rooftop can justify paying for a shopper who is nine months out, and because the same message serves every store in the market. Upper funnel, model-level, category-level, local events, the market-level content that answer engines quote. This is the spend that grows the denominator.

Demand capture belongs to the store, inside a defined trade area, with brand-term exclusivity by geography so the same branded click is not purchased twice. Store-level Performance Max, store-level inventory feeds, store-level offers. This is the spend that competes on service and price, which is what intra-brand competition is supposed to be about.

Co-op should follow incrementality, not participation. Pay against registrations in the trade area that would not have happened otherwise, not against the fact that the campaigns ran. A dealer council will accept that faster than it accepts a lecture, because it stops rewarding the neighbor for bidding up the local auction.

Measuring it without lying to yourself

The ratios that matter, in the order they become uncomfortable:

• Cost per incremental registration by trade area. Not per lead, not per form.
• Brand-term cost per click against category cost per click. When brand clicks cost more than category clicks, the market is bidding against itself.
• Unique reach duplication. How many of those impressions landed on the same households.
• Co-op dollars per incremental unit. The number that decides whether the program is a subsidy or a stimulus.

Then run the test that ends the argument. Hold one store's brand-term spend for thirty days, leave everything else in the market untouched, and watch two numbers: that store's registrations, and total same-brand registrations for the market. If the market total is flat while the paused store holds share, the spend was buying a sale that already existed. One matched geo, one month, one table. It travels better than any deck, and it is the same denominator discipline we apply to the campaign ladder in the performance report.

The short answer

Same-brand stores in one market should compete. That is the point of a franchise network, and a metro with one dominant rooftop is worse for everyone in it, including the brand. What should not happen is the brand paying all six to chase the identical shopper with identical campaigns and then reporting participation as growth.

Fund the demand nobody can pay for alone. Let the stores fight over capturing it. Measure what changed in the market, not what got bought in the auction.